Don't re-implement containing block logic when marking container chains for layout

Don't re-implement containing block logic when marking container chains for layout

When marking the container chain of an abspos object for layout we don't need to
worry about rel-pos inlines in the container chain ourselves - containingBlock()
does this for us. Since there is no situation in which we would want to mark
anything between an abspos object and its containing block with posChildNeedsLayout()
just use containingBlock() directly when climbing up the tree.

This is a refactoring supported by two tests that capture layout situations where
an abspos object has a container chain containing relpos inlines.

BUG=517369
Committed: https://crrev.com/440dd29c10e04558cd63853f4e9af2c8d182accc
Cr-Commit-Position: refs/heads/master@{#384247}

[rhogan, 2016-03-22 19:06:10]

robhogan@gmail.com changed reviewers:
+ mstensho@opera.com

[mstensho (USE GERRIT), 2016-03-29 10:56:29]

Need a blank line before the BUG= line. BTW: That bug was marked as fixed months
ago (with no patch, as far as I can tell; I don't remember the details), and the
code you're modifying is in the same area, but what's this change about, really?
Performance?

Both tests pass without your patch.

[rhogan, 2016-03-29 20:15:43]

Description was changed from

==========
Don't set needless layout on containers of positioned objects
BUG=517369
==========

to

==========
Don't re-implement containing block logic when marking container chains for
layout

When marking the container chain of an abspos object for layout we don't need to
worry about rel-pos inlines in the container chain ourselves - containingBlock()
does this for us. Since there is no situation in which we would want to mark
anything between an abspos object and its containing block with
posChildNeedsLayout()
just use containingBlock() directly when climbing up the tree.

This is a refactoring supported by two tests that capture layout situations
where
an abspos object has a container chain containing relpos inlines.

BUG=517369
==========

[rhogan, 2016-03-30 17:46:31]

On 2016/03/29 at 10:56:29, mstensho wrote:
> Need a blank line before the BUG= line. BTW: That bug was marked as fixed
months ago (with no patch, as far as I can tell; I don't remember the details),
and the code you're modifying is in the same area, but what's this change about,
really? Performance?
I've updated the description to clarify.

> 
> Both tests pass without your patch.
Yes, this is intended as refactoring.

[mstensho (USE GERRIT), 2016-03-31 11:12:59]

The CQ bit was checked by mstensho@opera.com

[mstensho (USE GERRIT), 2016-03-31 11:13:00]

lgtm

[commit-bot: I haz the power, 2016-03-31 11:13:15]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1323313004/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1323313004/100001

[commit-bot: I haz the power, 2016-03-31 11:56:14]

Committed patchset #6 (id:100001)

[commit-bot: I haz the power, 2016-03-31 11:57:18]

Description was changed from

==========
Don't re-implement containing block logic when marking container chains for
layout

When marking the container chain of an abspos object for layout we don't need to
worry about rel-pos inlines in the container chain ourselves - containingBlock()
does this for us. Since there is no situation in which we would want to mark
anything between an abspos object and its containing block with
posChildNeedsLayout()
just use containingBlock() directly when climbing up the tree.

This is a refactoring supported by two tests that capture layout situations
where
an abspos object has a container chain containing relpos inlines.

BUG=517369
==========

to

==========
Don't re-implement containing block logic when marking container chains for
layout

When marking the container chain of an abspos object for layout we don't need to
worry about rel-pos inlines in the container chain ourselves - containingBlock()
does this for us. Since there is no situation in which we would want to mark
anything between an abspos object and its containing block with
posChildNeedsLayout()
just use containingBlock() directly when climbing up the tree.

This is a refactoring supported by two tests that capture layout situations
where
an abspos object has a container chain containing relpos inlines.

BUG=517369

Committed: https://crrev.com/440dd29c10e04558cd63853f4e9af2c8d182accc
Cr-Commit-Position: refs/heads/master@{#384247}
==========

[commit-bot: I haz the power, 2016-03-31 11:57:19]

Patchset 6 (id:??) landed as
https://crrev.com/440dd29c10e04558cd63853f4e9af2c8d182accc
Cr-Commit-Position: refs/heads/master@{#384247}